Skip to main content
Premium Trial:

Request an Annual Quote

Stanford Team Devises PLA-Based Method for Single-Cell Measurement of RNA and Proteins


NEW YORK (GenomeWeb) – Stanford University researchers have developed a proximity ligation assay (PLA)-based approach for high-throughput simultaneous detection of RNA and protein in single cells.

Detailed in a paper published this week in Nature Methods, the approach, named PLAYR (proximity ligation assay for RNA), allows for measurement in the range of 40 to 50 RNAs and proteins in hundreds of thousands of cells, enabling a combination of multiplexing and throughput not previously available, Stanford's Garry Nolan, author on the paper and one of the method's developers, told GenomeWeb.

Key to the approach is use of PLA for RNA detection, which, Nolan said, enables higher multiplexing than methods like fluorescence in situ hybridization (FISH) and branched DNA, which are similarly compatible with high-throughput single-cell methods like flow cytometry and mass cytometry.

Originally developed by Uppsala University researcher Ulf Landegren, who Nolan noted is a frequent collaborator, PLA uses paired DNA probes to detect nucleic acids or, when attached to antibodies, proteins.

When the probes bind their targets (RNAs in the case of the assay presented in the Nature Methods study), the attached DNA probes are brought into proximity and ligate, forming a new DNA amplicon that can then be quantified. The quantity of the DNA corresponds to the quantity of the target.

Because the DNA labels can be constructed to hybridize only to their specific partner, PLA eliminates the problem of background binding, improving assay specificity. As Nolan noted, for signal to be generated, both probes have to be in close enough contact with each other to ligate, which is unlikely to happen unless they are bound to their intended target.

The primary challenge in devising the method was arriving at conditions that were appropriate for both the PLA-based RNA detection and for antibody-based protein detection.

"Getting this together with proteins at the same time really took quite a bit of effort and modification of the protocols and so on," Nolan said. "For example, if you don't prepare the cells correctly, the RNA gets degraded. On the other hand, what you do [to prepare the] RNA might denature the epitopes for the antibodies or cause the antibodies to fall off. So, there were a few steps and tweaks that need to be done to get it all working the right way."

Adding another wrinkle is the fact that, for best results, the researchers used multiple pairs of probes to each RNA target. This, Nolan said, ensured that if the binding site for one pair of probes was, for instance, blocked by a bound protein or the conformation of that particular RNA transcript, the other pairs of probes would potentially be able to bind and detect it.

"[In terms of] rules for how RNAs bind in cells or whether or not a protein is binding there, there is nothing in the sequence of the RNA that predicts that," he said. "So it is really just about hedging your bets."

In the paper, the authors recommended "using four to five probe pairs whenever possible" as "this number of probe pairs per gene generally led to reliable and robust detection of both rare and highly abundant transcripts."

This, though, requires careful selection of probes to make sure they would all work under the same conditions.

"We had to write a script that would make sure to double-check the sequences we were using against everything in the genome, make sure that the temperatures of annealing were comparable, things like that," Nolan said.

The approach can be used with a variety of platforms including microscope-based imaging, flow cytometry, and mass cytometry.

In the Nature Methods study, Nolan and his colleagues ran the approach on Fluidigm's CyTOF mass cytometry instrument, looking at 10 cell-surface proteins and their corresponding transcripts in primary human peripheral blood mononuclear cells, observing highly specific mRNA expression patterns between cell types as well as significant differences between protein and RNA levels.

Such a discrepancy between protein and RNA levels is not unexpected, having been observed in a number of previous experiments, Nolan said. The PLAYR data further hinted at the highly complex nature of this relationship, demonstrating not only that RNAs and proteins have different expression levels, but that these differences vary from cell to cell.

"Everyone has known that the assumption that RNA equals protein is a big assumption," Nolan said. "But we show that there is every possible pattern you could imagine waiting to be understood. There is protein with no RNA is some cells. There is RNA with no protein in other. There are obviously transcriptional and translational regulatory phenomena there waiting to be understood."

In addition to exploring differences between RNA and protein expression, the researchers used the approach to measure cytokines in different cell populations. While such measurements are typically made at the protein level, this has certain disadvantages, Nolan said.

Specifically, to measure cytokine expression, cells must be treated to prevent them from secreting these proteins, which can itself affect the biology of the cell. Additionally, these cytokine measurements can only be made hours after the cell is stimulated, making it impossible to collect certain other types of data such as on protein phosphorylation or transcription factor expression.

With the PLAYR approach, the researchers were able to measure cytokines at the RNA level, which Nolan said can be done 15 to 30 minutes after stimulation, while measuring proteins for other purposes such as distinguishing between different cell populations.

Nolan said that his lab recently submitted for publication another technique for simultaneous measurement of nucleic acids and proteins that will offer significantly higher multiplexing and throughput than the PLAYR method.

That method is capable of measuring tens of thousands of proteins and nucleic acids in millions of cells, he said, noting that it is based on an entirely different technology.